
浏览全部资源
扫码关注微信
1.上海理工大学大学能源与动力工程学院 上海 200093
2. TCL空调器(中山)有限公司 中山 528400
韩华,女,博士,副教授,上海理工大学能源与动力工程学院,13611880360,E-mail:happier_han@126.com。研究方向:制冷空调系统的故障诊断及优化,AI在制冷系统中的应用,高效机房与智慧运维。Han Hua, female, Ph. D., associate professor, School of Energy and Power Engineering, University of Shanghai for Science and Technology, 86-13611880360, E-mail: happier_han@126.com. Research fields: fault diagnosis and optimization of refrigeration and air conditioning system, application of AI in refrigeration system, high-efficiency plant room, smart operation & maintenance.
收稿:2025-09-17,
修回:2025-10-31,
录用:2025-11-21,
网络出版:2026-02-03,
移动端阅览
张宏宇,熊军,高旭等.基于物理信息引导半监督回归的小样本制冷剂量辨识[J].制冷学报,
Zhang Hongyu,Xiong Jun,Gao Xu,et al.Refrigerant Charge Identification with Limited Data via Physics-Guided Semi-Supervised Regression[J].Journal of Refrigeration,
张宏宇,熊军,高旭等.基于物理信息引导半监督回归的小样本制冷剂量辨识[J].制冷学报, DOI:10.12465/issn.0253-4339.20250917002. CSTR: XXXXX.XX.XXX.20250917002.
Zhang Hongyu,Xiong Jun,Gao Xu,et al.Refrigerant Charge Identification with Limited Data via Physics-Guided Semi-Supervised Regression[J].Journal of Refrigeration, DOI:10.12465/issn.0253-4339.20250917002. CSTR: XXXXX.XX.XXX.20250917002.
制冷剂泄漏是制冷空调系统能效损失的主要故障。本文创新性地提出融合物理信息引导的半监督回归模型(PG-SSR),将能量守恒方程计算的制冷剂循环量嵌入神经网络,作为模型的引导信息;并引入最大均值差异(MMD)与MC Dropout制冷剂量伪标签联合生成机制生成高置信度标签,利用大量未标记数据信息对训练样本进行扩容与增强,在小样本情景下实现制冷剂高精度定量预测。采用1台分体式空调制冷剂泄漏实验数据进行验证,结果表明,在同等小样本条件下,PG-SSR较基线模型的RMSE减少22.91 g,降幅达64.48%,MAPE减少3.86%,降幅达68.93%,性能显著优于基线模型,揭示了制冷剂循环量物理信息引导与联合校准的伪标签生成在制冷剂量辨识中的协同增益效果。
Refrigerant leakage is the dominant fault responsible for the degradation of energy efficiency in vapor compression systems. To address this issue, a physics-guided semisupervised regression (PG-SSR) model is proposed, in which the refrigerant circulation rate derived from the energy conservation equation embedded into the neural network as guiding information. In addition, a joint pseudo-label generation mechanism is introduced by combining the maximum mean discrepancy (MMD) with MC Dropout to produce high-confidence refrigerant charge labels. This strategy leverages large amounts of unlabeled data to augment and enhance limited training samples, thereby enabling an accurate quantitative prediction of refrigerant charges under small-sample scenarios. Validation usinf experimental data from a split-type air conditioner demonstrated that under identical limited-sample conditions, PG-SSR reduced the root mean square error (RMSE) by 22.91 g (64.48%) and mean absolute percentage error (MAPE) by 3.86% (68.93%) compared with baseline models. These results highlight the synergistic benefits of physical guidance and jointly calibrated pseudo-labelling in improving refrigerant charge identification performance.
Yang Liu , Yan Haiyan , Lam J C . Thermal comfort and building energy consumption implications-a review [J]. Applied Energy , 2014 , 115 : 164 - 173 .
陈志杰 , 朱旭 , 黄小清 , 等 . 基于混合模型的数据中心空调系统虚拟制冷剂充注量传感器 [J]. 制冷技术 , 2019 , 39 ( 6 ): 9 - 14 .
Chen Zhijie , Zhu Xu , Huang Xiaoqing , et al . Virtual refrigerant charge sensor of data center air conditioning system based on hybrid model [J]. Chinese Journal of Refrigeration Technology , 2019 , 39 ( 6 ): 9 - 14 .
Zhao Yang , Li Tingting , Zhang Xuejun , et al . Artificial intelligence-based fault detection and diagnosis methods for building energy systems: advantages, challenges and the future [J]. Renewable and Sustainable Energy Reviews , 2019 , 109 : 85 - 101 .
Guo Yabin , Liu Yaxin , Wang Yuhua , et al . Advance and prospect of machine learning based fault detection and diagnosis in air conditioning systems [J]. Renewable and Sustainable Energy Reviews , 2024 , 205 : 114853 .
Comstock M C , Braun J E , Groll E A . A survey of common faults for chillers [J]. ASHRAE Transactions , 2002 , 108 ( 1 ): 819 - 825 .
周志华 . 机器学习 [M]. 北京 : 清华大学出版社 , 2016 .
Zhou Zhihua . Machine learning [M]. Beijing : Tsinghua University Press , 2016 .
钟超文 , 花君 , 严珂 , 等 . 半监督支持向量机的空气处理机组夏季故障诊断 [J]. 中国计量大学学报 , 2018 , 29 ( 3 ): 311 - 316, 344 .
Zhong Chaowen , Hua Jun , Yan Ke , et al . Fault diagnosis for air handing units in summer seasons based on semi-supervised support vector machines [J]. Journal of China University of Metrology , 2018 , 29 ( 3 ): 311 - 316, 344 .
任正雄 , 韩华 , 崔晓钰 , 等 . 基于Tri-Training的制冷系统半监督故障诊断 [J]. 制冷学报 , 2022 , 43 ( 4 ): 129 - 136, 144 .
Ren Zhengxiong , Han Hua , Cui Xiaoyu , et al . Semi-supervised fault diagnosis of refrigeration system based on tri-training [J]. Journal of Refrigeration , 2022 , 43 ( 4 ): 129 - 136, 144 .
Lu Cheng , Ma Xiang , Yan Ke . Chiller fault diagnosis based on improved variational autoencoder and co-training framework: a case study of insufficient samples [J]. Journal of Building Engineering , 2024 , 88 : 109137 .
Li Haorong , Braun J E . Development, evaluation, and demonstration of a virtual refrigerant charge sensor [J]. HVAC&R Research , 2009 , 15 ( 1 ): 117 - 136 .
Kim W , Braun J E . Extension of a virtual refrigerant charge sensor [J]. International Journal of Refrigeration , 2015 , 55 : 224 - 235 .
Yang Yuting , Xu Ling , Han Hua , et al . Soft measurement and prediction of refrigerant leakage based on SVR-LSTM [J]. International Journal of Refrigeration , 2023 , 152 : 303 - 314 .
Jo Y , Kahng H , Kim S B . Deep semi-supervised regression via pseudo-label filtering and calibration [J]. Applied Soft Computing , 2024 , 161 : 111670 .
Ding Shifei , Zhang Zi , Xu Li , et al . A novel robust semi-supervised stochastic configuration network for regression tasks with noise [J]. Information Sciences , 2025 , 703 : 121933 .
Jiang Feifeng , Ma Jun , Li Zheng , et al . Prediction of energy use intensity of urban buildings using the semi-supervised deep learning model [J]. Energy , 2022 , 249 : 123631 .
Bu Xiangya , Wu Qiuwei , Zhou Bin , et al . Hybrid short-term load forecasting using CGAN with CNN and semi-supervised regression [J]. Applied Energy , 2023 , 338 : 120920 .
Wu Yuandi , Sicard B , Gadsden S A . Physics-informed machine learning: a comprehensive review on applications in anomaly detection and condition monitoring [J]. Expert Systems with Applications , 2024 , 255 : 124678 .
陈耀 , 韩华 , 储召平 , 等 . 基于物理引导神经网络的蒸发器制冷剂泄漏故障诊断模型 [J]. 暖通空调 , 2025 , 55 ( 8 ): 60 - 65 .
Chen Yao , Han Hua , Chu Zhaoping , et al . Fault diagnosis model of refrigerant leakage in evaporators based on physics-guided neural network [J]. Journal of HV&AC , 2025 , 55 ( 8 ): 60 - 65 .
Gretton A , Borgwardt K M , Rasch M J , et al . A kernel two-sample test [J]. Journal of Machine Learning Research , 2012 , 13 : 723 - 773 .
0
浏览量
0
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621